Sei sulla pagina 1di 6

Lim, Uzu, Takeuchi 1339

Lee Wah Lim Separation of benzene and deuterated benzenes by


Hideyuki Uzu
Toyohide Takeuchi reversed-phase and recycle liquid chromatography
using monolithic capillary columns
Department of Chemistry, Faculty
of Engineering, Gifu University,
1-1 Yanagido, Gifu 501-1193, An alternate pumping-recycle system utilizing a commercially available low dead-
Japan volume switching valve was developed for microcolumn LC. The recycle system had
two separation columns, and the dead volume of the recycling lines was kept to a
minimum by avoiding passage of the sample through the pump chamber, sample
injector, and the normal path length of a conventional UV detector. The drawback of
the high total back pressure caused by the second column that is placed after the
detector was overcome by on-column detection, and this eliminated the need for a
high pressure flow cell. The system was used for the separation of an authentic mix-
ture of benzene, benzene-1,3,5-d3, and benzene-d6. Baseline separation was accom-
plished after six cycles and the calculated theoretical plate number for benzene was
230,000. It was observed that the theoretical plate number (N) increased linearly with
increasing number of cycles, and the N per unit time increased with increasing inlet
pressure. The separation conditions were optimized and the separation of benzene
and benzene-d6 was accomplished within 75 min at 2.5 MPa inlet pressure.
Key Words: Recycle LC; Reversed-phase; Monolithic silica capillary column; Alternate-pumping;
Benzene; Deuterated benzene;
Received: July 13, 2004; revised: August 5, 2004; accepted: August 6, 2004
DOI 10.1002/jssc.200401882

1 Introduction employed [5 – 9]. In preparative chromatography, the


recycling technique is well known as a useful tool to
The task of obtaining high separation efficiency or high improve enantioseparation, production rate, and recovery
resolution in HPLC has been of great concern and been yield.
widely investigated over the past thirty years, as can be
seen from the evolution of the use of large spherical parti- Recycle chromatography (RC) is an approach in which
the sample undergoes recycling through the column(s),

Original Paper
cles to smaller ones in packed columns, from various
efforts made to increase the use of the ancillary tech- thus producing the same separation effect as is obtained
niques in micro LC [1 – 3], and from the recent introduction when a relatively long column is used. The advantages of
and development of a highly permeable monolithic sta- RC are that it allows the separation of compounds with
tionary phase [4]. Generally, two alternative methods are very close retention time, requires no increase in inlet
available when a difficult separation cannot be accom- pressure, and also eliminates any problems involved in
plished by changing the column and mobile phase condi- making long columns. On the other hand, once the ana-
tions. The most common approach is to increase the col- lytes are being separated, they would be mixed again dur-
umn length with identical packing material, but this might ing the recycling process. Thus peaks can only be
lead to a limitation of the separation efficiency due to an recycled until they begin to overlap.
increased inlet pressure. The second approach is to The concept of RC in LC has been studied since the late
recycle the sample through the column, which requires no 1960’s with the advent of HPLC. Basically, there are two
increase in the column inlet pressure. However, the dead operating modes in RC, i. e., direct pumping (DP) [10] and
volume in the recycle lines causes band broadening and alternate pumping (AP) [11 – 14]. In DP-RC, a pump, a
peaks can only be recycled until they begin to overlap. sample injector, a column, and a detector are connected
Thus, the recycling technique has been limited to prepara- in series in a closed loop, as shown in Figure 1 (DP). The
tive chromatography where larger diameter columns are effluent from the detector is fed back to the pump, and the
unresolved sample is recycled within the loop until suffi-
Correspondence: Toyohide Takeuchi, Department of Chemis-
try, Faculty of Engineering, Gifu University, 1-1 Yanagido, Gifu cient resolution is attained, whereas in AP-RC the pump is
501-1193, Japan. Phone/Fax: +81 58 293 2806. removed from the recycle loop and the sample is recycled
E-mail: takeuchi@apchem.gifu-u.ac.jp. exclusively through the column(s). Figure 1 (AP1) and

J. Sep. Sci. 2004, 27, 1339 – 1344 www.jss-journal.de i 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1340 Lim, Uzu, Takeuchi

mode [17 – 19]. Cartoni and Ferretti [17] reported the ear-
liest separation of benzene-deuterobenzene and toluene-
deuterotoluene with reversed-phase LC, and separation
factors higher than those of GC were obtained. For all the
compounds tested in refs. [17 – 19], the deuterated com-
pounds always eluted before their hydrogen analogs.
Baweja [18] suggested that this chromatographic beha-
vior depends fundamentally on the interaction between
the C – H or C – D bonds and the stationary phase. A C – H
bond has a higher oscillation frequency (3300 cm – 1) than
a C – D bond (2334 cm – 1), and it induces greater forces of
attraction (i.e., the van der Waals forces) between itself
and the stationary phase, thus leading to longer retention.
This is supported by the results obtained by Turowski et
al. [19] when they examined the reversed-phase chroma-
tographic separation of 10 pairs of isotopic compounds on
10 different stationary phases having alkyl and aryl groups
bonded to the silica surface.
Sj. van der Wal has studied the isotopic separation of ben-
zene and deuterated benzenes by reversed-phase micro-
bore and recycle HPLC [20] and the limitations of high-
efficiency separation by alternate column RC [21]. By
using a 225 cm60.32 mm fused silica column packed
with 5 lm reversed-phase material, the separation of ben-
zene-d6, monodeuterobenzene, and benzene was
achieved at approximately 680 min with a 15% aqueous
acetonitrile (ACN) eluent under 40 MPa inlet pressure.
This separation was improved with a recycle system using
two conventional 15 cm64.6 mm columns packed with
3 lm material, the separation of monodeuterobenzene
and benzene was achieved after 19 cycles at
approx. 454 min with a 40% aqueous ACN eluent under
Figure 1. Simplified schematic diagrams of the DP and AP 37 MPa inlet pressure [20]. By considering the limiting
recycle designs. I: 4-port valve. II: 6-port valve. III: 8-port effects of retention factors, such as the increased inlet
valve. See text for details (Section 1). pressure, a virtual baseline separation of monodeutero-
benzene and benzene was accomplished by generating
200,000 apparent theoretical plates with two 15-cm col-
(AP2) shows the AP recycle systems that utilize a 6-
umns per cycle after 55 cycles in approximately 12 h
port [13] or an 8-port [14] switching valve, respectively.
under 38 MPa and 20% aqueous ACN elution [21].
The main advantage of the 8-port design over the 6-port
design is that only one detector is needed to monitor the The present paper introduces the use of highly permeable
sample. However, the 8-port design possesses a trivial monolithic silica capillary columns in a newly designed AP
disadvantage, i.e., the direction of flow through the detec- recycle system for microcolumn LC in order to increase
tor is reversed every time the valve is switched. In general, the column efficiency. One of the main advantages of the
there are five main criteria that need to be taken into proposed AP system is that only one detector is used to
account when choosing the mode of operation, i. e., monitor the transfer of analytes from one column to the
(1) dead volume of the pump, (2) cost of the column, other given the same flow direction. Compared to the use
(3) availability of a high-pressure flow cell on the detector, of conventional packed columns in RC reported so far, the
(4) instrumentation control such as the necessity to use of monolithic capillary columns is expected to achieve
actuate the switching valve at regular intervals, and better isotopic separations in terms of the retention time,
(5) solvent consumption. the elution profile, the drawback of too high inlet and back
pressure, and the optimum number of cycles. The system
The separation of partially or completely substituted isoto- was easily assembled with the aid of a commercially avail-
pic molecules with deuterium atoms has been done able low dead-volume 6-port switching valve, and the iso-
mainly by GC [15, 16] and by LC in the reversed-phase topic separation of deuterobenzene, 1,3,5-trideuteroben-

J. Sep. Sci. 2004, 27, 1339 – 1344 www.jss-journal.de i 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Separation of benzene and deuterated benzenes by RP and RC LC 1341

zene, and benzene by reversed-phase LC was investi- 2.2 Reagents


gated. Under optimal conditions, a virtual baseline sep- HPLC-grade ACN and distilled water were obtained from
aration was achieved after 6 cycles in approximately 6 h Nacalai Tesque (Kyoto, Japan). Special grade benzene
at a low inlet pressure, i.e., 0.8 MPa. was purchased from Wako Pure Chemical Industry
(Osaka, Japan), while laboratory use benzene-1,3,5-d3
2 Experimental and benzene-d6 were obtained from Aldrich Chemical Co.
(Milwaukee, WI, USA). Tetramethoxysilane (TMOS) and
2.1 Apparatus
tetrachlorosilane were purchased from Tokyo Chemical
The recycle system, as shown in Figure 2, consisted of a Industry (Tokyo, Japan). Poly(ethylene glycol) with
nitrogen gas-driven pumping system, a Model 7000 10,000 average molecular weight (PEG 10,000) was
switching valve (Rheodyne, Cotati, CA, USA), a M435 obtained from Aldrich (USA). Other reagents were of
micro injection valve (Upchurch Scientific, Oak Harbor, guaranteed grade and were obtained from Nacalai Tes-
WA, USA), two laboratory-made monolithic columns que. All reagents were used without any further treatment,
(length ranging from 215 to 540 mm60.10 mm ID), a unless otherwise noted. The ratio of ACN to water in the
Model CE-1570 Intelligent UV/Vis detector (Jasco, Tokyo, eluent was optimized for the isotopic separation.
Japan), and a Chromatopac C-R7Ae plus data processor
(Shimadzu, Kyoto, Japan). The pressure of the nitrogen
supplied from a gas cylinder was regulated by a model 2.3 Monolithic capillary columns
CFC14PP flow controller (Shimadzu). A PTFE tube with a
Monolithic silica capillary columns were prepared as pre-
total volume of 0.8 mL, which was used as the eluent loop,
viously reported [2, 4] with some modifications in the pre-
was connected via the Rheodyne switching valve. Accord-
paration conditions. A 2-mL volume of TMOS and 0.53 g
ing to our experimental results (unpublished observa-
PEG 10,000 were dissolved in 0.01 M acetic acid by agi-
tions), the dead volume of the Upchurch switching valve is
tating the solution for 30 min at 08C. The solution was
0.05 lL.
degassed for 10 min before it was filled into the pretreated
0.10 mm ID fused-silica capillaries (GL Sciences, Tokyo,
Japan). After leaving the capillaries at 408C for 24 h, they
were washed with water and 0.1 M aqueous ammonia
solution, and were kept at 608C for 45 h. The capillaries
were heated at 3308C for 5 h after washing with 60% etha-
nol, and then purged with nitrogen gas at 1108C for 1 h.
The capillaries were then reacted with 10% dimethylocta-
decylchlorosilane in toluene solvent at 1408C for 20 h, fol-
lowed by washing with toluene, tetrahydrofuran, metha-
nol, and the eluent prior to use.

2.4 Alternate-pumping (AP) procedure


In the AP-RC developed in this study, the pump was
removed from the recycle loop and the sample was
recycled exclusively through two monolithic silica capillary
Figure 2. Schematic diagram of the AP recycle system for
monolithic capillary columns. columns, which were connected in tandem to a six-port
two-position switching valve (Upchurch Scientific). As
shown in Figure 3, when the switching valve was at posi-
tion A, the isotopic compounds eluted from column 1 to
The CE-1570 UV detector was used for on-column detec- column 2. Once the signals were detected when the com-
tion. A ca. 1 cm length of the polyimide coating of the col- pounds passed through column 2, the valve was switched
umn that was connected to the detector was peeled off by to position B in order to direct the isotopic compounds
a razor and was fixed to the cell block of the detector. back onto column 1 from column 2. By calculating the
Sample injection was performed via the sample injection retention of the compounds on both columns, the valve
port. During the injection of the sample, the flow of the elu- was switched to position A before the compounds left col-
ent was stopped by switching the upper valve (Rheodyne) umn 1 to prevent them being drained out. By manually
to the middle position, and then switching it back to its ori- switching the valve during these intervals, the compounds
ginal position, i. e., the load position, as soon as the sam- were recycled exclusively through the two columns until
ple is being loaded. sufficient resolution is attained.

J. Sep. Sci. 2004, 27, 1339 – 1344 www.jss-journal.de i 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1342 Lim, Uzu, Takeuchi

Figure 4. Recycle separation of benzene-d6 and benzene by


on-column detection. Column: monolithic silica ODS, 540
Figure 3. Illustration of the eluent flow path when the switch- and 410 mm60.10 mm ID. Mobile phase: ACN-water
ing valve is set at position A or position B. See text for oper- (40 : 60). Inlet pressure: 1 MPa. Sample: 0.1% for each C6D6
ating details (Section 2.4). 1 & 2: monolithic capillary column and C6H6. Wavelength of UV detection: 208 nm.
1 and 2 (0.10 mm ID).
In effect, the developed recycle design led the isotopic
compounds from column 1 fi 2 fi 1 fi 2 fi 1 and so forth
3 Results and discussion until sufficient resolution is attained. Since the actual
length of the columns is constant, there is no increase in
3.1 Significance of the AP recycle system the system back-pressure. Compared with other recycling
The main aim of this work was to develop a high resolution techniques, the band broadening of the compounds was
recycle system for microcolumn LC using monolithic capil- kept to a minimum as they do not recycle through the
,
lary columns in order to attain a high separation efficiency pump and the sample injector as well as the detector’ as
as if a relatively long column were being used. The resolu- on-column detection was applied. With on-column detec-
tion, Rs, of two chromatographic peaks is defined as: tion, the monolithic capillary columns could be connected
pffiffiffiffi directly through the cell block of the detector and the
Rs = [ N /2] [(a – 1)/( a + 1)] [kAV/(1 + kAV)] (1) switching valve without any extra connection tubes, thus
eliminating the need for a high pressure flow cell. How-
where N is the number of theoretical plates, a is the sep-
ever, due to the limited path length as well as the pre-
aration factor, and kAV is the average retention
pffiffiffiffi factor of sence of the stationary phase, this might lead to poor sen-
the two peaks. Thus, Rs is proportional to N and hence
sitivity.
to l, the length of the column. So a 4-fold increase in the
length of the column would double the resolution, and this
is accompanied by an increase in the retention time. The 3.2 Separation of benzene and deuterated
recycling technique, which practically increases the length benzenes
of the column, would theoretically increase the resolution The system was then used for the separation of an
of a separation. authentic mixture of benzene, benzene-1,3,5-d3, and ben-
Figure 4 shows the separation of benzene-d6 and ben- zene-d6, as shown in Figure 5. For Figure 5.A, operating
conditions as in Figure 4 were applied, and the sample
zene via the recycle system. A virtual baseline separation
was prepared by using ACN-water (30 : 70) solution. The
was accomplished by generating 150,000 and 160,000
separation of the isotopic molecules was unsatisfactory
plates for both benzene-d6 and benzene, respectively,
even after nine cycles. By reducing the strength of the elu-
with on-column detection. With a total length of 95 cm per
ent from ACN-water (40 : 60) to ACN-water (30 : 70), and
cycle, the isotopic separation was achieved after 5 cycles
in approximately 5 h at 1 MPa and by using 40% aqueous the inlet pressure from 1 to 0.8 MPa, a baseline separa-
ACN eluent with UV detection at 208 nm. The N calcu- tion was attained after six cycles and the calculated N for
lated for the first detected benzene-d6 peak was 30,000 benzene as shown in Figure 5.B was 230,000.
and this number multiplied as the compound recycled
through the columns. After the second, third, fourth, and 3.3 Optimization of the operating conditions
the fifth cycle, the N for benzene-d6 were 60,000, 90,000, The relationship between the inlet pressure and N as well
120,000, and 150,000, respectively. as the linear flow rate for the recycle system was investi-

J. Sep. Sci. 2004, 27, 1339 – 1344 www.jss-journal.de i 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Separation of benzene and deuterated benzenes by RP and RC LC 1343

Table 1. The effects of inlet pressure and the theoretical plate number per unit time in recycle chromatography. Operation condi-
tions as in Figure 6.
Inlet pressure Theoretical plate Theoretical plate Linear flow N per unit
[kPa] number [N] height [lm] rate [mm s – 1] time [min – 1]
400 28400 25.0 0.326 403
500 29100 24.4 0.404 510
600 30500 23.3 0.478 635
700 28800 24.7 0.553 692
800 27800 25.6 0.594 715
900 26200 27.1 0.720 819
1,000 25100 28.3 0.795 881
1,500 23100 30.7 1.157 1176
2,000 21200 33.5 1.565 1458
2,500 19600 36.3 1.940 1674
3,000 16900 42.0 2.316 1727
3,500 15300 46.2 2.733 1850

Figure 6. Theoretical plate number per unit time vs. applied


inlet pressure in recycle chromatography. Column: mono-
lithic silica ODS, 495 and 215 mm60.10 mm ID. Mobile
phase: ACN-water (40 : 60). Sample: C6H6. Wavelength of
UV detection: 208 nm.

obtained with related applied inlet pressure, which ranged


from 0.4 to 3.5 MPa, and with ACN-water (40:60) used as
the mobile phase, were as summarized in Table 1. By
increasing the inlet pressure, the N decreased accompa-
nied by an increase in theoretical plate height. However,
the N per unit time increased with increasing applied inlet
pressure. Figure 6 shows that further increase of the inlet
pressure beyond 2.5 MPa did not improve the N per unit
Figure 5. Recycle separation of benzene-d6 (1), benzene-
1,3,5-d3 (2), and benzene (3) with on-column detection. (A): time significantly but only increased the burden applied to
Operating conditions as in Figure 3, except for the sample. the separation columns.
Sample: 0.1% for each C6D6, C6D3H3 and C6H6 in ACN-water
(30 : 70) solution. (B): Column: monolithic silica ODS, 455 By considering all these parameters, the separation con-
and 440 mm60.10 mm ID. Mobile phase: ACN-water ditions were optimized and the isotopic separation of ben-
(30 : 70). Inlet pressure: 0.8 MPa. Sample: same as (A).
Wavelength of UV detection: 208 nm. zene-d6 and benzene was accomplished within 75 min
after 5 cycles at 2.5 MPa and with 40% aqueous ACN
gated. On using a monolithic capillary column with a total used as the mobile phase, as shown in Figure 7. Within a
column length of 71 cm and benzene as the analyte, the N much shorter analysis time, higher N values were attained

J. Sep. Sci. 2004, 27, 1339 – 1344 www.jss-journal.de i 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1344 Lim, Uzu, Takeuchi

ing higher separation efficiency. In other words, any diffi-


cult separation could possibly be achieved by recycling
the sample through the columns until sufficient resolution
is attained and this actually reduces the turnaround time
as optimization of the separation method on a conven-
tional chromatographic system can be very time consum-
ing.

Acknowledgement
The authors gratefully acknowledge the Takeda Founda-
tion for financial support.

Figure 7. Recycle separation of benzene-d6 (1) and ben-


References
zene (2) under optimized conditions. Column: monolithic [1] L.W. Lim, T. Miwa, T. Takeuchi, Anal. Sci. 2001, 17 Supple-
silica ODS, 495 and 470 mm60.10 mm ID. Mobile phase: ment, i887 – i890.
ACN-water (40 : 60). Inlet pressure: 2.5 MPa. Sample: C6D6 [2] T. Takeuchi, S. Tatsumi, S. Masuoka, K. Hirose, H. Uzu,
and C6H6 in ACN-water (30 : 70) solution. Wavelength of UV J.-Y. Jin, C. Fujimoto, K. Ohta, K.-P. Lee, J.-J. Ryoo, S.-H.
detection: 208 nm. Choi, J. Chromatogr. A 2003, 1021, 55 – 59.
[3] L.W. Lim, K. Hirose, S. Tatsumi, H. Uzu, M. Mizukami, T.
Takeuchi, J. Chromatogr. A 2004, 1033, 205 – 212.
compared to that found under the earlier conditions (as in [4] F. Svec, T.B. Tennikova, Z. Deyl, Monolithic Materials: Pre-
Figure 4). At the fifth cycle, the N for both benzene-d6 and paration, Properties and Applications. Elsevier, Amsterdam
benzene were 120,000, respectively. 2003.
[5] T. Yoshida, C.-K. Shu, E.T. Theimer, J. Chromatogr. 1977,
137, 461 – 464.
[6] W.M.P. Johnson, D.F. O’Keefe, K. Rihs, J. Chromatogr.
4 Conclusions 1984, 291, 449 – 452.
[7] T. Okano, S. Masuda, S. Kusunose, M. Komatsu, T.
The developed AP recycle system is suitable for the sep- Kobayashi, J. Chromatogr. 1984, 294, 460 – 465.
aration of isotopic molecules, either partially or completely [8] J.R. Crary, K. Cain-Janicki, R. Wijayaratne, J. Chromatogr.
substituted with deuterium atoms. Even though the sys- 1989, 462, 85 – 94.
tem requires two identical separation columns, the dead [9] W.S. Letter, J. Chromatogr. 1992, 590, 169 – 173.
volume of the recycling lines was kept to a minimum by [10] K.J. Bombaugh, W.A. Dark, R.F. Levangie, J. Chromatogr.
Sci. 1969, 7, 42 – 47.
avoiding passage of the sample through the pump cham-
[11] J.A. Biesenberger, M. Tan, I. Duvdevani, T. Maurer, J. Poly-
ber, sample injector, and the normal path length of a con-
mer Sci. 1971, B9, 353 – 357.
ventional UV detector. Since the standard flow cells on [12] I. Duvdevani, J.A. Biesenberger, M. Tan, J. Polymer Sci.
most UV absorption detectors are incompatible with the 1971, B9, 429 – 434.
total back pressure caused by the second column that is [13] R.A. Henry, S.H. Byrne, D.R. Hudson, J. Chromatogr. Sci.
placed after the detector, high-pressure flow cells are 1974, 12, 197 – 199.
needed. This drawback was overcome by using a system [14] K. Lan, J.W. Jorgenson, Anal. Chem. 1998, 70, 2773 – 2782.
that utilizes two monolithic silica capillary columns con- [15] G.P. Cartoni, A. Liberti, A. Pela, Anal. Chem. 1967, 39(13),
1618 – 1622.
nected in tandem, and with on-column detection. It was
[16] M. Possanzini, A. Pela, A. Liberty, G.P. Cartoni, J. Chroma-
observed that N increased linearly with increasing number togr. 1968, 38, 492 – 497.
of cycles and the N per unit time improved with increasing [17] G.P. Cartoni, I. Ferretti, J. Chromatogr. 1976, 122, 287 –
applied inlet pressure. It was also noted that the deuter- 291.
ated forms eluted before their hydrogen analogs. [18] R. Baweja, Anal. Chim. Acta 1987, 192, 345 – 348.
[19] M. Turowski, N. Yamakawa, J. Meller, K. Kimata, T. Ike-
For difficult isotopic separations that require long analysis gami, K. Hosoya, N. Tanaka, E.R. Thornton, J. Am. Chem.
times, the recycle system utilizing capillary columns Soc. 2003, 125, 13836 – 13849.
reduces the consumption of mobile phase. RC also pro- [20] Sj. Van der Wal, J. Liquid Chromatogr. 1985, 8, 2003 – 2016.
vides a relatively simple and inexpensive mean of attain- [21] Sj. Van der Wal, Chromatographia 1986, 22, 81 – 87.

J. Sep. Sci. 2004, 27, 1339 – 1344 www.jss-journal.de i 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Potrebbero piacerti anche